Early postnatal lung development in the eastern quoll (Dasyurus viverrinus)

Early postnatal lung development (1–25 days) in the eastern quoll (Dasyurus viverrinus) was investigated to assess the morphofunctional status of one of the most immature marsupial neonates. Lung volume, surface density, surface area, and parenchymal and nonparenchymal volume proportions were determined using light microscopic morphometry. The lungs of the neonate were at the canalicular stage and consisted of two “balloon‐like” airways with few septal ridges. The absolute volume of the lung was only 0.0009 cm3 with an air space surface density of 108.83 cm−1 and a surface area of 0.082 cm2. The increase in lung volume in the first three postnatal days was mainly due to airspace expansion. The rapid postnatal development of the lung was indicated by an increase in the septal proportion of the parenchyma around day 4, which was reflected by an increase in the airspace surface density and surface area. By day 5, the lung entered the saccular stage of development with a reduction in septal thickness, expansion of the tubules into saccules and development of a double capillary system. The subsequent saccular period was characterized by repetitive septation steps, which increased the number of airway generations. The lungs of the newborn Dasyurus viverrinus must be considered as structurally and quantitatively insufficient to meet the respiratory requirements at birth. Hence, cutaneous gas exchange might be crucial for the first three postnatal days. The lung has to mature rapidly in the early postnatal period to support the increased metabolic requirements of the developing young.


| INTRODUCTION
Mammalian lung development is generally characterized by three major periods: embryonic, fetal, and postnatal. The phases of lung development are based primarily on morphological criteria: the embryonal period encompasses lung organogenesis; the fetal lung development comprises pseudoglandular, canalicular, and saccular stages; and finally, postnatal development comprises alveolarization and microvascular maturation (Schittny, 2017).
Recently, an increasing number of studies on lung development have been reported for marsupials. This is because marsupials offer a unique model for understanding mammalian lung development. Marsupials are distinguished from eutherian mammals by their extremely small size and marked immaturity at birth (Figure 1; Ferner et al., 2009;Ferner, Schultz, & Zeller, 2017). Most of their development and growth, including that of the lung, occurs postnatally and are supported by a prolonged lactation period (Ferner & Mess, 2011;Renfree, 2006). Hughes and Hall (1988) sorted marsupial neonates into three grades of developmental complexity (G1, G2, and G3) based on size variation and the developmental degree of their organ systems. These grades correspond more or less to the altricial-precocial-spectrum of eutherians: G1 (Dasyuridae) is the least developed, G2 (Peramelidae, Phalangeridae, and Didelphidae), and G3 (Macropodidae and Phascolarctidae) represent the most developed marsupial neonates ( Figure 1). However, marsupial neonates are always more immature than neonates of altricial eutherians.
The lung structure of marsupial neonates follows the size variation in sequences G1 to G3. Gradation of lung development from the canalicular stage to the saccular stage can be observed in newborn marsupials (Ferner, 2018;Mess & Ferner, 2010).
The respiratory system in marsupials was generally considered adequately developed prior to birth, to a level where it was capable of gas exchange commensurate to the metabolic needs of the neonate (Tyndale-Biscoe & Janssens, 1988). However, increasing evidence suggests that the immaturity of the respiratory system at birth in the marsupial necessitates the recruitment of an alternative organ system, such as the skin for gas exchange (Frappell & MacFarlane, 2006;MacFarlane & Frappell, 2001). Given the relative immaturity at birth and the large surface area to volume ratio inherent with small body size, it is not surprising that cutaneous gas exchange occurs to varying extents during the neonatal period of the marsupial (Ferner, 2018;Simpson et al., 2011).
Gas exchange in newborns of two dunnart species was conducted solely via cutaneous respiration (95%-100%) for a number of days; no thoracic movement or pulmonary ventilation was observed, and a rhythmic breathing pattern was not apparent until day 4 (Frappell & MacFarlane, 2006;Frappell & Mortola, 2000;Mortola, Frappell, & Woolley, 1999;Simpson et al., 2011). These physiological studies indicate that lung maturation must proceed very quickly in dasyurids to meet the respiratory requirements of the developing young.
Dasyurid marsupials offer a unique opportunity for a better understanding of mammalian lung development. In contrast to eutherian mammals, the canalicular stage of lung development is encountered postnatally, and the entire process of postnatal lung development occurs in a ventilated functioning state.
The current study used histological studies to examine the structural development and possible functional capacity of the maturing lung during the postnatal period of a very small newborn marsupial, the eastern quoll (Dasyurus viverrinus), whose young are born after 19 days of gestation and with a birth weight of just 12 mg.

| Animal provenance
A total of 45 eastern quoll (Dasyurus viverrinus) pouch youngs between neonate (<24 hr) and 25 days, from which 29 specimens were examined histologically, were enrolled in this study. The specimens were available from the Hubrecht & Hill collection, which is a part of the embryological collection of the Museum für Naturkunde, Leibniz-Institut für Evolutions-und Biodiversitätsforschung, Berlin.
The Hill collection encompasses a large ontogenetic series of pre-and postnatal stages of eastern quolls, which were collected in Australia between 1898 and 1906 by James Peter Hill (Klima & Bangma, 1987). The collection is dedicated to research and has already been used in numerous scientific studies. These specimens were serially sectioned and whole-mount fixed samples. Details about fixation and processing of the material are unknown, but it is most probable that the material was fixed with Bouin's solution and preserved in 70% ethanol as state-of-the-art at this time. The fixative Bouin's solution (picric acid, acetic acid, and formaldehyde in an aqueous solution) and the entire embedding process in paraffin for histology can cause shrinkage of the material (~30%, see Ross, 1953). Considering this shrinkage during the processing of the histological material, the obtained absolute values of lung volume (V L ) and surface area (S a ) might be underestimated, but comparable to each other because of the same treatment of the material. However, the morphometric measurements of the volume densities of lung components are unaffected by volume changes because these are volume fractions.
The age, head length (HL), and crown-rump length (CRL) of the specimens were documented, and a staging system for the development of D. viverrinus (A-M) was provided by James Peter Hill. To estimate body weight, 16 alcohol-preserved specimens of the corresponding ontogenetic stages of D. viverrinus were investigated. Body weights were matched to the serially sectioned specimens based on the crown-rump length and head length values. The body weights and specifics and numbers of the specimens used in this study are summarized in Table 1. The number of animals investigated per age stage varied depending on the availability in the Hubrecht & Hill collection. Morphological development of the ontogenetic stages of D. viverrinus, investigated here for lung development (0-25 days), is presented in Figure 2.

| Lung morphometry
Serial histological sections of whole embedded specimens were investigated by light microscopy using a stereomicroscope (Zeiss Axiokop, Carl Zeiss Microscopy GmbH, Germany) equipped with a digital camera (Leica DFC490, Leica Microsystems, Switzerland Ltd.), connected to a computer (software LAS V4.2, Leica Microsystems, Switzerland Ltd.). Systematic sampling for estimation of the volume fractions of the lung components was combined with the Cavalieri volume estimation to obtain the absolute volume. The entire right and left lungs were examined for lung morphometry, and the stereological analysis followed the procedure of systematic uniform random sampling (SURS) as proposed by Hsia, Hyde, Ochs, and Weibel (2010). To ensure that the selected lung sections represent the whole and all parts of the lung have an equal probability of being sampled, the fractionator method was applied. The lungs were serially sliced to a constant thickness (mostly 10 μm). The total length of the lung was calculated from the number of all lung sections, and a fractionator sequence with a sampling fraction of 1/8 was applied. The first section selected for morphometry was randomly chosen using a random number table within the sampling fraction. Thus, for each specimen, eight micrographs were obtained at magnifications of ×50 or ×100 depending on the size of the lung. The morphometric analysis program STEPanizer (Tschanz, Burri, & Weibel, 2011) was used to determine the lung volume and volume densities of parenchymal and nonparenchymal lung components.
Lung volume was determined as a reference to translate relative stereological parameters (e.g., volume and surface density) into absolute values by using the Cavalieri principle and point counting methods (Hsia et al., 2010;Tschanz, Schneider, & Knudsen, 2014). The total lung volume was calculated as follows: F I G U R E 2 Postnatal development of Dasyurus viverrinus. The macrographs reflect the morphological transformation from neonate to day 25: neonate (a) 0-12-hr-old, (b) 7-day-old, (c) 14-day-old, (d) 19-day-old, and (e) 25-day-old. The highly immature neonate (a) had a massive oral shield and large prominent nostrils; eye and ear primordia were barely visible and a definitive neck was missing. The characteristics of marsupial neonates are strongly developed forelimbs and rudimentary hindlimbs. On day 7 (b), a definitive neck was visible; the nasal swelling was moderate and a simple oral shield was present. The morphological transformation of the fore-and hindlimbs started at this time. A clear developmental progress occurred between days 14 (c) and 19 (d), where ear and eye primordia and the area of the later mystacial vibrissae became visible. On day 25 (e), the head appeared even more differentiated, with a lateral lip line, eyelids, and auricle, which were still fused. Scale bar = 0.5 cm where P is the number of points falling on the lung tissue, d is the distance between adjacent grid points, and t is the distance between the selected lung sections.
To examine the structural changes in the developing lung, the volume densities of the lung parenchyma (Vvp) and nonparenchyma (Vvnp) were obtained by point counting methods as proposed by Makanya, Haenni, and Burri (2003). Furthermore, the volume densities of the coarse constituent components of the parenchyma, tubular, or saccular air spaces (Vva), and septa (Vvs), were estimated.
The volume densities of the nonparenchymal components were also estimated: conducting airways (Vvb), blood vessels with diameters >25 μm (Vvv), and connective tissue (Vvt). The magnification for the morphometric analysis of the lung structure was ×100-×200. All points on the parenchyma and those on the nonparenchymal and its respective components were scored using the STEPanizer computer program (Tschanz et al., 2011). The morphometric measurements followed the multistage stratified analysis (Hsia et al., 2010).
The volume of the parenchyma Vvp, for example, was calculated from the equations: Vvp = Pp*Pt − 1 and Vp = Vvp*V L where V L is the volume of the lung, and Pp and Pt are the point counts on the parenchyma and the total number of points on the reference space, respectively.
The air space surface area Sv(a) of the lungs was obtained using the intersection counting method (see Howard & Reed, 2005 for details). Multiplying Sv(a) by the volume density of the lung parenchyma and the absolute lung volume yielded the absolute surface area S(a).

| Statistical analysis
All the age group data are presented as mean ± 1 SD. Differences in the weight-specific lung volumes and volume proportions of the various components of the lung were statistically tested using one-way analysis of variance (ANOVA). Age groups (days 0, 2, 3, 4, 5, and 14) that had statistically viable numbers of animals were tested. Tukey's test was used to identify the specific values with differences. The data in the form of ratios were arcsine-transformed before the statistical analysis was performed. All values were set at a significance level of p < .05.

| Lung structure
The structural changes in the lung during the first 25 postnatal days are summarized in Figures 3, 4, 5, and 6. The whole lung of D. viverrinus comprised a single left lung and a right lung with an accessory lobe (Figure 3b). The left lung lobe was smaller than the right lobe. Fissures subdividing the lung were not evident in the lungs of the neonate but developed by days 4-5 (Figure 4e,g). The left and right lobes were each served by the main bronchus or emerged directly from the trachea for the first postnatal day (Figures 3e,f, and 4a). The accessory lobe was intimately joined to the right lobe through the parenchymatous lung tissue (Figure 3b and 4b).
The lungs of the neonate D. viverrinus were at the canalicular stage of lung development, characterized by primitive "balloon-like" airways that consisted of large tubules. Only a few septal ridges protruded into the lumen and poorly subdivided the lung (Figure 3a-c). Some respiratory capillaries were located at the inner surface of the lungs, suggesting that gas exchange might be possible. The respiratory epithelium is located on the opposite side of the lung wall from the visceral epithelium. The inner surface of the lungs was lined mainly with the cuboidal epithelium. Capillaries with abundant nucleated erythrocytes abutted onto the epithelium in a few areas, forming the first gas exchange barrier with thin type I squamous epithelium (Figure 3d).
On day 2 of age, the septal ridges can be seen growing, leading to further subdivision of the lung (Figure 3g). The lung was still at the canalicular stage, and the tubules were predominately lined with a cuboidal epithelium with interspersed respiratory capillaries ( Figure 3h). By day 3, the development of the bronchial tree had started. The short main bronchi opened immediately into the tubular airways of the right and left lobes, connecting the accessory lobe to the right lobe (Figure 4a,b). The septation of the lung continued, and the newly formed intertubular septa were rather thick. Although the cuboidal epithelium still dominated the lining of the tubular airways, many respiratory capillaries were present and alternated with the parenchymal epithelium ( Figure 4c).
The septation of the lung progressed during the following days, and the pulmonary vascular system developed (Figure 4d, e). By day 4, the lungs appeared well subdivided, with several tubular structures. The septa between the tubules resembled double capillary septa, indicating a transition from the canalicular to the saccular stage. However, a consistent double-capillary network was not evident because of persistent small stretches of the cuboidal epithelium (Figure 4f). By day 5, the lung F I G U R E 3 Histological sections of the lung in the (a-d) neonate and (e-h) 2-day-old Dasyurus viverrinus. The lung of the newborn D. viverrinus consisted of poorly vascularized airspaces (tubules). The respiratory epithelium is located on the opposite side of the wall of the lung from the visceral epithelium. The epithelial lining of the airways consisted of mostly cuboidal cells, interspersed with a few respiratory capillaries (d, inset). No double capillary septa were present. This points to the canalicular stage of lung development. The lung appeared mostly undivided, and only a few septal ridges were protruding from the walls. (g) The lung of the 2-day-old joey resembled that of the neonate; however, more septal ridges subdivided the lung. A bronchial tree was not present yet, and the airspaces opened directly from the trachea (e, f). The epithelium lining the tubules were mainly cuboidal with some respiratory capillaries in between (h, inset). a, accessory lobe; aa, ascending aorta; ce, cuboidal epithelium; dea; descending aorta; h, heart; i, intestine; k, kidney (mesonephros); l, liver; ll, left lung; o, oesophagus; r, rib; rc, respiratory capillary; re, respiratory epithelium; revc, right vena cava; rl, right lung; sc, spinal cord; sr, septal ridge; t, trachea; tu, tubule; ve, visceral epithelium; vb, vertebral body. The respiratory capillaries are indicated by arrowheads. The magnification is indicated by the scale bar F I G U R E 4 Histological sections of the lung in the (a-c) 3-day-old, (d-f) 4-day-old, and (g-h) 5-day-old Dasyurus viverrinus. By day 3, the lung became more compartmentalized by thick septal ridges (a). Thick intertubular septa were lined by parenchymal epithelium alternating with respiratory capillaries (c, inset). The bronchial system consisted of short main bronchi that opened immediately to the tubular airways. By day 4, the sub septation of the lung continued and numerous smaller tubules resulted, a double capillary system was developing, indicating the transition from the canalicular to the saccular stage of lung development (f, inset). In the 5-day-old young, the sub septation of the lung parenchyma progressed resulting in numerous large saccules. (a), accessory lobe; (b), bronchus; dea; descending aorta; h, heart; ll, left lung; lmb, left main bronchus; mb, main bronchus; o, oesophagus; pa, pulmonary artery; pv, pulmonary vein; r, rib; rc, respiratory capillary; revc, right vena cava; rl, right lung; rmb, right main bronchus; sa, saccule; sr, septal ridge; t, trachea; tu, tubule; vb, vertebral body. The respiratory capillaries are indicated by arrowheads. The magnification is indicated by the scale bar had entered the saccular stage of development with a reduction in septal thickness, expansion of the tubules into saccules, and development of a consistent double capillary system (Figure 4g,h).
During the following 2 days, the lung development progressed and led to further subdivision of the lung parenchyma (Figure 5a,e). By day 6, the bronchial trees began to form. Several short smooth-walled conducting airways (later lobar bronchi) branched off the main bronchi and communicated directly with the large terminal saccules (Figure 5a-c). The large saccules were characterized by numerous septal ridges, leading to the formation of new saccules (Figure 5c,f). The thick septa separating the smaller saccules consisted of a double capillary network and an abundant cellular and acellular connective tissue layer in between (Figure 5d,g-j). Centrally located capillaries or small blood vessels were often encountered in the septa (Figure 5d,h,j). Large blood vessels were commonly found at the thick septal junctions (Figure 5h). By day 7, the bronchial tree consisted of the main and lobar bronchi. The smooth-walled conducting airways extended to the lung periphery, where they opened into the terminal saccules (Figure 5e). The larger saccules in the lung of the 7-day-old D. viverrinus were due to numerous septal ridges more irregular in shape than those seen in earlier stages (Figure 5f), whereas the smaller new-formed saccules were smooth-walled ( Figure 5h). Septa with a double capillary network separated the numerous small saccules.
Between days 14 and 25, the parenchymal structure increased in complexity (Figure 6). The development of the bronchial tree progressed, and lobar and segmental bronchi were present (Figure 6a,b,d,e,g,h). The lung parenchyma was further subdivided into numerous smaller saccules separated by thick septa. At 25 days, the air compartments of the lung parenchyma had markedly increased in number and decreased in size (compare Figures 3bf;4b,e,g;5b,f;and 6a,d,h). By day 14, 19, and 25, many of the respiratory capillaries were positioned close to the airspaces on both sides of a thickened interstitium, indicating the presence of a double capillary network (Figure 6c,f,i). By day 25, the double capillary septum appeared slightly thinner due to a reduction in the interstitial layer.
3.2 | Lung morphometry 3.2.1 | Lung volume, surface densities, surface area, and body mass Body mass and lung volumes (V L ) for the various developmental stages of D. viverrinus are presented in Table 2 and Figure 7.
At birth, the neonate weighed approximately 0.01 g and the lung volume was only 0.000894 ± 0.000348 cm 3 . By day 2, both bodyweight and lung volume nearly doubled. From day 3 to day 25, a steady increase in lung volume was observed (Figure 7). Specific V L was significantly higher in neonates and 2-day-old young than in all other developmental stages. Overall the developmental stages, V L was closely correlated with body mass (r = .986).
The surface densities Sv(a) and absolute surface areas S(a) for the airspaces are listed in Table 2. The airspace surface density was lowest in the neonate D. viverrinus (108.83 19.93 cm −1 ). The airspace surface density increased steadily with progressive structural lung development via subseptation of the lung parenchyma. Sv(a) more than doubled from day 1 to day 25 (260.53 cm −1 ). The absolute surface area of the neonatal lung (0.082 ± 0.020 cm 2 ) increased 26-fold by day 14 (2.122 ± 0.569 cm 2 ). The airspace surface area was positively correlated with the body mass ( Figure 7). When scaled against V L , S a increased at a rate close to that of the lung volume ( Figure 8).

| Lung composition
The volume densities of the various coarse parameters of the lung are provided in Table 3, and the absolute volumes are scaled against the body mass in Figure 9.
The densities of the parenchyma (Vvp) ranged from 0.779 ± 0.014 (day 7) to 0.900 ± 0.019 (day 1); the highest value was observed in the neonatal group. Within the parenchymal components, the volume density of the airspaces Vva was highest in the neonatal group (0.627 ± 0.081) and remained significantly higher than the proportion of the septal tissue (Vvs) for the first three postnatal days (> 0.5). From days 4 to 25, the volume density of the airspaces decreased and remained low. The volume density of the parenchymal septa was reciprocal to that of the alveolar space. During the first 3 days, the volume densities of the septa were low and increased to higher values from days 4 to 25. A shift around day 4 was clearly noticeable when comparing the absolute volumes of parenchymal components (Figure 9). The volumes of both parenchymal components increased with body mass; however, the volume of the septal tissue was higher than that of the air spaces.
The values for the nonparenchymal volume density were reciprocal to the volume density of the parenchyma, with a low of 0.100 ± 0.019 by day 0 and a high of 0.221 ± 0.014 by day 7. The volume density of the nonparenchymal showed a slight (not significant) increase from days 5 to 6 on. The primary airways and the F I G U R E 5 Histological sections of the lung in the (a-d) 6-day-old and (e-j) 7-day-old Dasyurus viverrinus. By day 6, the development of the bronchial tree had progressed. Several short bronchi (lobar bronchi), branching off the main bronchi, communicated directly with large terminal saccules. The large saccules were further subdivided by septal ridges. The thick septa separating the smaller saccules consisted of a double capillary network (d, inset). By day 7, three generations of bronchi were present. The bronchial tree consisted of the main, lobar, and segmental bronchi, which communicated with the terminal saccules. A septum with a double capillary network separated the numerous small saccules (j, inset). b, bronchus; bv, blood vessel; dea; descending aorta; la, left atrium; ll, left lung; lmb, left main bronchus; mb, main bronchus; o, oesophagus; pa, pulmonary artery; pv, pulmonary vein; r, rib; ra, right atrium; rc, respiratory capillary; rl, right lung; rmb, right main bronchus; s, septum; sa, saccule; sr, septal ridge. The respiratory capillaries are indicated by arrowheads. The magnification is indicated by the scale bar pulmonary vascular system began to develop around this time. The values of the volume densities of the nonparenchymal components (airways, blood vessels, connective tissue) are summarized in Table 3, and the absolute values are presented in Figure 9. The differences were not statistically significant. The volume densities of the airways (Vvb) and blood vessels (Vvv) were lowest in the neonatal group and highest by day 7. Generally, the volume densities and absolute volumes of the airways, blood vessels, and connective tissue showed a tendency to increase during the saccular period, reflecting the formation of the bronchial tree and pulmonary vasculature.

| Lung structure
It can be assumed that mammalian lung development is highly conserved and follows similar developmental pathways in all mammalian species, including marsupials and monotremes (Ferner et al., 2009;Schittny, 2017;Szdzuy et al., 2008). However, difficulties in the classification of marsupial lung development may arise from the fact that mammalian lung development has been studied mainly in eutherian species and definitions for the different stages of lung development (e.g., canalicular, saccular, and alveolar) focused on the situation found in altricial eutherian species, mostly during the prenatal period in the nonventilated state. In marsupials, the characteristics of lung developmental stages might be slightly different, because lung maturation takes place postnatally in a ventilated and functioning state. Compared to eutherians, marsupials and monotremes are born earlier during the developmental trajectory of lung development. Generally, the lungs of newborn marsupials are reported to be in the terminal sac (saccular) stage (Gemmell & Nelson, 1988;Krause & Leeson, 1975;Mess & Ferner, 2010;Runciman et al., 1996Runciman et al., , 1998 or even the canalicular stage of lung development (Burri, Haenni, Tschanz, & Makanya, 2003;Ferner, 2018;Makanya et al., 2003;Makanya, Sparrow, Warui, Mwangi, & Burri, 2001;Makanya, Tschanz, Haenni, & Burri, 2007;Modepalli et al., 2018;Simpson et al., 2011).
However, all marsupial newborns show qualitative characteristics of a mature gas-exchanging organ, such as the full complement of surfactant proteins (Makanya et al., 2007;Miller, Orgeig, Daniels, & Baudinette, 2001;Ribbons, Baudinette, & McMurchie, 1989) and a thin blood-gas barrier (Runciman et al., 1996;Szdzuy et al., 2008). This seems to be valid even for the most immature marsupial neonates, as reported for the newborn dasyurids Dasyurus hallucatus (Gemmell & Nelson, 1988) and Sminthopsis crassicaudata (Simpson et al., 2011). In contrast, branching morphogenesis, which is completed during the canalicular stage in fetal eutherians, is postponed in marsupials and takes place postnatally during the saccular period. This indicates that lung structures indispensable for respiratory function (e.g., stabilization of air spaces by surfactant and thin blood-gas barrier for gas exchange) have to be mature in newborn marsupials, whereas the formation of distal airways is delayed and follows the structural development of the lung parenchyma.
The first description of the lung of the newborn eastern native cat, D. viverrinus, originated from Hill and Hill (1955). They stated that the lungs are remarkably simple organs, being structurally at the lowest level, compatible with some degree of functional efficiency. A later ultrastructural study of the lungs of the newborn northern native cat D. hallucatus described the lung as similar in structure to two balloons, which were lined internally with a respiratory epithelium formed by both squamous cells and surfactant-secreting cells (Gemmell & Nelson, 1988). This describes the lung structure of the newborn D. viverrinus quite well and is confirmed by the present study. However, it appears difficult to fit this "balloon-like" lung structure into the developmental trajectory of mammalian lung development. The parenchymal phenotypes during pre-and postnatal lung development have been well documented and used, among other characteristics, to define the different stages of lung development (Burri, 1999;Schittny, 2017). Previous descriptions of the canalicular stage were based on species that were not exposed to air. However, the presence of stretches of cuboidal epithelium and a few portions of the thin blood-gas barrier in the inner lining of the lung, as observed in the newborn D. viverrinus, are some of the essential characteristics of the canalicular stage. The inner lining of the air spaces resembles that described in the lungs of the neonates of the northern native cat (Gemmell & Nelson, 1988) and the quokka (Makanya et al., 2007), two marsupial species that also have lungs at the canalicular stage at birth. Gemmell and Nelson (1988) stated that the respiratory lining of the lung is composed of surfactant-secreting cells (cuboidal) and squamous cells that were separated from the endothelial cells of the underlying capillaries by a composite of two basal laminae; thus, a true blood-gas barrier was present. However, Makanya et al. (2007) suggested that the squamous epithelium differentiates earlier than the capillary system. The delay in the formation of capillaries and their opposition to the tubular epithelium indicates increased diffusion distance in several parenchymal regions (Makanya et al., 2007). F I G U R E 6 Histological sections of the lung in the (a-c) 14-day-old, (d-f) 19-day-old, and (g-i) 25-day-old Dasyurus viverrinus. By day 14, the lung parenchyma appeared further subdivided with numerous saccules separated by thick double capillary septa (c, inset). Until 25 days, a reduction in the size of the saccules and an increase in the surface area available for gas exchange were achieved by septation. By day 25, the double capillary septum appeared thinner due to a reduction in the interstitium (i, inset). br, bronchiolus; h, heart; lb, lobar bronchus; mb, main bronchus; o, oesophagus; pa, pulmonary artery; r, rib; rc, respiratory capillary; s, septum; sa, saccule; sb, segmental bronchus; sr, septal ridge. The respiratory capillaries are indicated by arrowheads. The magnification is indicated by the scale bar In the most mature marsupial neonates (G3), a functional lung develops prenatally in a very short time. The formation of a functional lung occurs in the bandicoot in the last 3-4 days of gestation (Gemmell & Little, 1982), in which the canalicular stage is converted into the saccular stage in the last 3 days of gestation (Runciman et al., 1996). Marsupial species, born with lungs at the canalicular stage, attain the saccular stage rapidly after birth. In the quokka wallaby, for example, the canalicular stage is converted to the saccular stage within the first four postnatal days (Makanya et al., 2001). In addition, the gray short-tailed opossum and the fat-tailed dunnart reached the saccular stage in the postnatal period by days 8 and 10, respectively (Modepalli et al., 2018;Simpson et al., 2011). In the lungs of the D. viverrinus examined in this study, the formation of the primary septa started by day 6, indicating that the transition from canalicular to saccular stage occurred around this time. The following saccular period was rather long, and the lung was still at the saccular stage on day 25, when this study ended. The saccular stage was characterized by the formation of transitory saccules, which were progressively subdivided by septation into more generations of saccules. The process of saccule multiplication is very similar to that of alveolar formation (Burri, 1974) and is accompanied by tissue proliferation . In contrast to alveolization, microvascular maturation, a process that leads to the  Note: The body mass used for calculation of the specific lung volume was measured in alcohol preserved eastern quoll specimens with corresponding age and body size (see Table 1). "*" value significantly different (p ≤ .05, ANOVA, Tukey). Values are given as means ± SD for those age groups where more than one animal was measured, n is the number of individuals in the group. Since the subsequent course of lung development could not be followed in D. viverrinus, the time of alveolization remains unknown. However, the structural lung development of D. viverrinus may follow a time course similar to that of the fat-tailed dunnart, a dasyurid species with similar birth weight (13 mg) and developmental degree at birth (G1). Simpson et al. (2011) reported that in the fat-tailed dunnart, the first secondary septa did not occur until day 45 when alveolization started. The period between days 70 and 100 was the most important time for bulk alveolization (Simpson et al., 2011).

| Lung volume and body mass
At birth, D. viverrinus weighs approximately 12 mg and has a lung volume (V L ) of only 0.000894 ml. Bodyweights of 750 g in females and 1.1 kg in males have been reported for adult eastern quolls (Cooper & Withers, 2010). This means that from birth to adulthood, the bodyweight increase lies in the range of 63,000-92,000 times. If the V L increase is proportional to body weight, the V L of adult D. viverrinus may be approximately 56-82 ml. However, the actual lung volume might be lower, since comparison to marsupials with similar body mass revealed lower V L values (e.g., quokka wallaby (210-day-old, mass: 711 g, V L : 21.3 mL; Burri et al., 2003) and tammar wallaby (250-day-old, mass: 775 g, V L : 30.5 ml; Runciman et al., 1998).
Since the present study examined lung development only in the early postnatal period (until day 25), the course of complete lung development cannot be described for D. viverrinus. Given the fact that V L doubled in 2 days, it can be assumed that in D. viverrinus, the Values are given as means ± SD for those age groups where more than one animal was measured, n is the number of individuals in the group.
F I G U R E 9 Bilogarithmic plots of the volumes of the air spaces (Va), the septal tissue (Vs), the conducting airways (Vb), the coarse blood vessels (Vv), and the connective tissue (Vt) against body mass for Dasyurus viverrinus in the postnatal period. The graphs are based on group means, and the regression lines are provided most dramatic increase in V L occurs within the first postnatal stage of development (~3-5 postnatal days) when the lung changes from the canalicular stage to the saccular stage. The increase in V L may be mainly due to airspace expansion, as indicated by the significantly high volume densities of the airspaces during the first three postnatal days. It has been reported that V L increase from birth to adulthood is approximately 23 times in humans and rats Zeltner, Caduff, Gehr, Pfenninger, & Burri, 1987), 3,800 times in the tammar wallaby (Runciman et al., 1998), and 8,000 times in the quokka wallaby . Considering the enormous gain in body mass from newborns (12 mg) to adults (780-1,100 g) in D. viverrinus, the increase in V L should be even more dramatic, possibly approximately 30,000 times. However, the low absolute V L in the neonate D. viverrinus may be related not only to the small body size, but also to the stage of lung development. Makanya et al. (2001) proposed that the earlier the developmental stage of lung development, the lower the V L at birth.

| Surface densities and surface area
The absolute surface area of the airspace increased 89-fold from day 1 to day 25, reflecting a remarkable increase in the surface area for gas exchange. The massnormalized surface area of the newborn D. viverrinus (0.82 m 2 kg −1 ) was low compared to the mass-normalized values of airspace surface area reported for other newborn marsupials (quokka wallaby: 1.34 m 2 kg −1 , tammar wallaby: 1.22 m 2 kg −1 ), and for altricial eutherians (newborn rat: 5.31 m 2 kg −1 ; Makanya et al., 2007).
The low surface density of the air spaces in the newborn lung reflected the simple "balloon-like" structure at birth. Marsupial newborns with developmental degrees 2 and 3 have more subdivided lungs than newborn dasyurids, which is reflected by the higher surface densities of the air spaces (for review see Ferner, 2018).
Continuous reorganization of the septal components during the early postnatal period (days 1-25) of Dasyurus viverrinus resulted in a doubling of the surface density of the air spaces.

| Lung composition
The proportion of the lung parenchyma Vvp (90%) in the newborn D. viverrinus is comparable to that reported for the quokka wallaby (93%; Makanya et al., 2007) and the tammar wallaby (88%; Runciman et al., 1998). The low proportion of nonparenchymal (10%; conducting airways, blood vessels, and connective tissue) in the lungs of the newborn D. viverrinus reflects the virtual nonexistence of a bronchial tree and poor vascularization at birth.
The high proportion of the airspaces (63%) in the newborn lung of D. viverrinus result from the "balloonlike" lung structure with only a few septal ridges protruding in the airspace. The proportions of airspaces reported for the newborn quokka wallaby (70%; Makanya et al., 2007) and tammar wallaby (63%, Runciman et al., 1998) are quite similar.
Starting from day 4, a reversal in the parenchymal proportions (air spaces, 36%; septa, 50%), reflected the extensive septal formation and increasing subdivision of the lung parenchyma, which continued in the saccular period. A similar situation was observed in the 3-day-old quokka wallaby, where the proportions of the airways and septa were 30% and 58%, respectively .
Since the present study examined the lungs of D. viverrinus only in the early postnatal period, when the lung was still at the saccular stage, no statements can be made about the timing of alveolization and possible changes in lung composition during this period. However, this process might be similar to that reported for other marsupial species. In the quokka wallaby, after a relatively brief canalicular stage, a prolonged saccular stage markedly increased the lung parenchyma. The terminal wave of saccule septation, in the quokka around day 125, led to the formation of the first generation of alveoli, causing a dramatic increase in the gas exchange area . It can be hypothesized that the process of alveolization might occur earlier in D. viverrinus than in the quokka wallaby. However, the structural changes in the lungs associated with alveolization might be comparable.

| Gas exchange through the skin
In the neonate of D. viverrinus, the small lung volume and the highly immature lung structure, resulting in low airspace surface density and total surface area, might affect pulmonary function. Although the lungs of the newborn D. viverrinus show qualitative characteristics (surfactant system and blood-gas-barrier) of a gasexchanging organ, quantitatively, this organ is poorly developed to meet the metabolic demands at birth. It is most probable that the neonate of D. viverrinus is almost totally dependent on the gas exchange via the skin, as has been reported for other newborn dasyurids (Mortola et al., 1999;Frappell & MacFarlane, 2006).
Recent studies on the skin structure of D. viverrinus revealed that the neonate possesses an extensive subepidermal capillary network, characterized by low diffusion distances and high capillary volume density, a welldeveloped vascular system for communication between cutaneous capillaries and the cardiac system, and an undivided ventricle (Ferner, 2018(Ferner, , 2020. These structural prerequisites allow for extensive transcutaneous gas exchange. The duration of cutaneous gas exchange during the postnatal period seems to be determined primarily by the maturation of the cardiorespiratory system. In particular, the closing of the shunts, resulting in the separation of the left and right ventricles, necessitates a transition from cutaneous to pulmonary gas exchange around 3 days after birth (Ferner, 2020;Runciman et al., 1995). The rapid postnatal development of the lung in D. viverrinus is indicated by a marked increase in the septal proportion of the parenchyma around day 4, which is reflected by an increase in the air space surface density and surface area at this time. These findings support the assumption that the time of transition from cutaneous to pulmonary respiration was around day 4. The pulmonary system has to mature quickly to be functional to meet the metabolic needs of the developing young.

| CONCLUSION
Generally, a functioning lung at birth must have a large surface area for gas exchange, a thin blood-gas barrier, a surfactant system, a conductive airway tree, appropriately developed vasculature, and coordinated neuromuscular effort. Despite the presence of several qualitative characteristics that allow for gas exchange, for example, surfactant system and blood-gas barrier, other important characteristics, such as large surface area, bronchial tree, vascular system, and neuromuscular control are poorly developed or missing entirely in the newborn D. viverrinus. Thus, structurally and quantitatively, the lungs of the newborn D. viverrinus must be considered insufficiently developed to permit the gas exchange to meet the metabolic needs at birth. Therefore, it is most probable that D. viverrinus, a marsupial species with one of the smallest neonates, must rely on cutaneous gas exchange for the first three postnatal days, and this is made possible only by a low metabolic rate and favorable body surface area. In the early postnatal period, the lungs of D. viverrinus mature rapidly, characterized by extensive tissue proliferation and septation, leading to an increase in gas exchange surface area, to meet the metabolic requirements of the developing young.